Research On Credibility Model And Evaluation Method Of Online Shopping Product Reviews Based On Knowledge Graph

With the rapid development of online shopping,people are more enthusiastic about purchasing goods on major online e-commerce platforms,and various evaluations of goods have become a key factor in consumer shopping decisions.The product evaluation of the e-commerce platform contains a lot of useful information for consumers and merchants,and these massive evaluations often hide some useless information and misleading information,which greatly reduces the credibility of the reviews.Based on online product reviews,this article studies the design of the credibility of the reviews and the research on the evaluation methods.It aims to provide new solutions for consumer shopping decisions and the supervision of reviews by industry and commerce departments.The knowledge graph is a structured semantic knowledge base,which can describe the concepts and mutual relations between objects in a simple symbol form.The knowledge contained in the reviews is composed of many concepts,including high-level concepts with rich semantics and low-level concepts with single semantics.Based on the improved knowledge graph,this article gives weight to the knowledge in product reviews,through reasonable design The credibility model can achieve the effect of quantifying the credibility of the comment.This article has conducted in-depth research on the design of credibility models and the evaluation methods.The main work is as follows:(1)This article extract the relevant entities and attributes of mobile phones from professional mobile phone evaluation websites and domain expert knowledge,and build mobile phone knowledge graphs based on existing knowledge graph construction methods.At the same time,this paper expands mobile phone knowledge graphs to give relevant knowledge weight,which lay the foundation for subsequent credibility model design.(2)Based on the designed mobile phone knowledge graph,this paper designs the weighted product knowledge graph by expanding the mobile phone knowledge graph,and extract the knowledge with higher similarity in the relevant reviews and the knowledge graph to construct the review semantic web.Finally,a credibility model for online shopping product reviews is designed based on the review semantic network(3)This paper studies the function of the evaluation credibility measurement model on the review's credibility ranking of e-commerce platforms.And we verify the feasibility and effectiveness of the model by comparing with the Jingdong recommendation model and other researchers' ranking model.The evaluation of the model is carried out with the help of questionnaires and indicators such as recall rate and accuracy rate.